AlphaGeometry Conquers Olympiad-Level Geometry

<p>Fermat&rsquo;s Last Theorem, scribbled in the margin of an ancient Greek math text in 1637 by Pierre de Fermat, went unsolved for centuries. The&nbsp;<em>Guinness Book of World Records</em>&nbsp;called it the &ldquo;most difficult mathematical problem&rdquo; due to the number of unsuccessful attempts, down through the centuries, to prove it. It was solved in 1994 by British mathematician Andrew Wiles, an event that landed Wiles half a dozen of the most prestigious prizes available to mathematicians, as well as a building named after him at the University of Oxford.</p> <p>Is this something that could be automated? This is the question that occurred to&nbsp;<a href="https://research.google/people/trieu-h-trinh/" rel="noopener ugc nofollow" target="_blank">Trieu H. Trinh</a>, a recent NYU Computer Science PhD graduate, at the outset of his PhD four years ago. Wiles&rsquo; achievement was impressive, but 357 years was a really long time to wait for a solution. &ldquo;What if we could build an AI that could do the same?&rdquo; he asked, &ldquo;so that next time if we want to solve a hard problem, we don&rsquo;t have to wait hundreds of years?&rdquo;</p> <p><a href="https://nyudatascience.medium.com/alphageometry-conquers-olympiad-level-geometry-2dfcd4709dd6"><strong>Read More</strong></a></p>